High school involvement and student engagement as protective factors for at-risk students

There continue to be a large number of at-risk students who do not complete high school every year. There are a number of identifiable risk factors that can contribute to an increased likelihood of students dropping out of high school. With advances in data collection, schools are now better able to identify and track students' progress towards graduation with detection systems, called Early Warning Systems (EWS). EWS utilize data on grades, behavior referrals, and attendance gathered from school records to identify students at increased risk for dropout. Students identified by schools as "at-risk" or "off-track" can then be provided with effective interventions designed to prevent dropout. Student engagement is one variable that schools have the ability to measure and potentially increase through interventions. EWS can be used to help facilitate linking "at-risk" and "off-track" students, who potentially report low school engagement, to a school's preexisting intervention programs in order to prevent dropout. Furthermore, participation in extracurricular activities provided by the school may help make students feel more connected and engaged at school. This can be particularly important for students transitioning from middle school to high school. These transition programs, set up to help connect the incoming class with upper classmates, are a great way for students to acclimate to the high school setting. With the different programs in place within a high school, it is important that students are connected with the programs and services that are right for them to help facilitate engagement and connectedness to school. Ensuring engagement and connectedness to school can positively impact grades, attendance, and behavior, and also decrease the likelihood of dropping out. The current study aimed to confirm the model that participation in at-risk programs has a positive impact on student engagement, which in turn, positively impacts student outcomes, such as grades, attendance, and behavior. The study found that participation in at-risk programs did not necessarily improve school outcomes or student engagement; however, students within these programs who reported higher school engagement had better school outcomes.

HIGH SCHOOL INVOLVEMENT AND STUDENT ENGAGEMENT AS PROTECTIVE FACTORS FOR AT-RISK STUDENTS by Chloe Ruebeck A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Educational Psychology The University of Utah December 2015 Copyright © Chloe Ruebeck 2015 All Rights Reserved The Uni v e r s i t y o f Ut ah Gr a dua t e S choo l STATEMENT OF DISSERTATION APPROVAL The dissertation of Chloe Ruebeck has been approved by the following supervisory committee members: Daniel Olympia , Chair 9/28/15 Date Approved Lora Tuesday Heathfield , Member 9/28/15 Date Approved John Kircher , Member 9/28/15 Date Approved Jason Burrow-Sanchez , Member 9/28/15 Date Approved Hollie Pettersson , Member 9/28/15 Date Approved and by Anne Cook , Chair/Dean of the Department/College/School of Educational Psychology and by David B. Kieda, Dean of The Graduate School. ABSTRACT There continue to be a large number of at-risk students who do not complete high school every year. There are a number of identifiable risk factors that can contribute to an increased likelihood of students dropping out of high school. With advances in data collection, schools are now better able to identify and track students' progress towards graduation with detection systems, called Early Warning Systems (EWS). EWS utilize data on grades, behavior referrals, and attendance gathered from school records to identify students at increased risk for dropout. Students identified by schools as "at-risk" or "off-track" can then be provided with effective interventions designed to prevent dropout. Student engagement is one variable that schools have the ability to measure and potentially increase through interventions. EWS can be used to help facilitate linking "at-risk" and "off-track" students, who potentially report low school engagement, to a school's preexisting intervention programs in order to prevent dropout. Furthermore, participation in extracurricular activities provided by the school may help make students feel more connected and engaged at school. This can be particularly important for students transitioning from middle school to high school. These transition programs, set up to help connect the incoming class with upper classmates, are a great way for students to acclimate to the high school setting. With the different programs in place within a high school, it is important that iv students are connected with the programs and services that are right for them to help facilitate engagement and connectedness to school. Ensuring engagement and connectedness to school can positively impact grades, attendance, and behavior, and also decrease the likelihood of dropping out. The current study aimed to confirm the model that participation in at-risk programs has a positive impact on student engagement, which in turn, positively impacts student outcomes, such as grades, attendance, and behavior. The study found that participation in at-risk programs did not necessarily improve school outcomes or student engagement; however, students within these programs who reported higher school engagement had better school outcomes. This dissertation is dedicated to my amazing family who provided support and encouragement throughout my journey as a graduate student. TABLE OF CONTENTS ABSTRACT ....................................................................................................................... iii LIST OF FIGURES ......................................................................................................... viii LIST OF TABLES ............................................................................................................. ix ACKNOWLEDGEMENTS ............................................................................................... xi CHAPTERS 1. INTRODUCTION ...................................................................................................1 United States' Dropout Statistics and Trends ..........................................................3 Negative Impact of Dropping Out ...........................................................................8 Theories Behind High School Dropout..................................................................11 Student Engagement ..............................................................................................14 Risk Factors for Dropout .......................................................................................21 Early Warning Systems..........................................................................................27 Prevention and Intervention Programs ...................................................................32 Strategies Specific to the Ninth-Grade Year ..........................................................36 The Impact of Extracurricular Activities ...............................................................40 Utah Dropout Trends and Prevention Initiatives ...................................................43 Rationale for Current Study ...................................................................................46 Research Questions ................................................................................................48 2. METHOD ..............................................................................................................52 Participants .............................................................................................................52 Setting ....................................................................................................................54 Measures ................................................................................................................59 Procedures ..............................................................................................................65 Study Design ..........................................................................................................67 3. RESULTS ..............................................................................................................80 Descriptive Statistics ..............................................................................................80 Inferential Statistics ...............................................................................................80 vii 4. DISCUSSION ......................................................................................................113 Main Findings ......................................................................................................114 Secondary Research Questions ...........................................................................123 Strengths and Limitations ....................................................................................130 Implications for Future Research and Practice ....................................................134 APPENDICES A. STUDENT ENGAGEMENT INSTRUMENT ...............................................139 B. STUDENT SURVEY ......................................................................................142 REFERENCES ................................................................................................................147 LIST OF FIGURES Figures 1 Structural Model for At-risk Participation ......................................................................73 2 Main At-Risk Structural Model ......................................................................................74 3 Main Prevention Structural Model..................................................................................75 4 Conditional Latent Growth Curve Analysis: Days Absent .............................................76 5 Conditional Latent Growth Curve Analysis: ODRs .......................................................77 6 Conditional Latent Growth Curve Analysis: GPA .........................................................78 7 Secondary Research Questions 6 and 7 Structural Model ..............................................79 LIST OF TABLES Tables 1 Demographic Characteristics of Sample 1 (All Grades).................................................70 2 Demographic Characteristics of Sample 2 (9th and 10th Grades) ..................................72 3 AMOS Variables ...........................................................................................................105 4 Fit Indices for Research Question 1 ..............................................................................106 5 Regression Weights: Research Question 1 Using Model 2 ..........................................106 6 Fit Indices for Main At-Risk and Prevention Model ....................................................106 7 Regression Weights: Main At-Risk Model 2 ................................................................107 8 Total Effects: Main At-Risk Model 2 ...........................................................................107 9 Post-hoc Regression Weights: Main Prevention Model 3 ............................................108 10 Post-hoc Total Effects: Main Prevention Model 3 .....................................................108 11 Fit Indices for Secondary Research Question 5 ..........................................................109 12 Regression Weights: Secondary Research Question 5 ...............................................109 13 Fit Indices for Secondary Research Questions 6 and 7 ...............................................110 14 Means and Standard Error for Student Engagement Variables by Grade ..................110 15 Means and Standard Error for EWI Variables ............................................................110 16 Research Question 8 Fit Indices .................................................................................110 17 Secondary Research Question 8: Race .......................................................................111 18 Secondary Research Question 8: Grade ......................................................................111 x 19 Secondary Research Question 8: SES.........................................................................111 20 Secondary Research Question 8: Gender ....................................................................112 ACKNOWLEDGEMENTS I would like to start of by thanking my dedicated chair, Lora Tuesday Heathfield. I could not have completed my dissertation without her guidance, knowledge, support, and patience. I would also like to thank the rest of my dissertation committee: Dan Olympia for his unwavering support and kind words; John Kircher for his incredible statistics knowledge, his time and effort helping me analyze my data, and for his patience as I better understood Structural Equation Modeling techniques and programs; Jason Burrow-Sánchez for his knowledge and guidance in better understanding Structural Equation Modeling techniques and programs; Hollie Pettersson for her encouragement, guidance, knowledge, and for connecting me with different educational resources and stakeholders. I would also like to thank Canyons School District and Jordan High School for allowing me to complete my study. I would especially like to thank Tom Sherwood for his understanding, and for providing me time and computers to complete my study. I would also like to thank my family and friends. They have been an incredible support throughout my dissertation, and graduate school career. I am truly lucky to have them. CHAPTER 1 INTRODUCTION Many school districts in the United States are continuing to take steps to positively impact graduation rates. This is in part due to legislation that places significant pressure on schools to increase their graduation rates and improve student outcomes. With more of an emphasis on outcomes, researchers have focused on using student outcome data to drive school policy and programs. However, even with significant gains in data-based decision making and improved policy and programs, there continue to be a number of students who disengage and leave school systems each year without earning a high school diploma or equivalent. The students who leave school are at a significant disadvantage compared to their peers who graduate from high school. The students who drop out are more likely to struggle to find employment, to be on public assistance, use drugs, be arrested, and spend time in prison (Dynarski & Gleason, 2002). In order to continue to improve graduation rates, research needs to continue in many promising areas. One area of research that has gained momentum is the study of student engagement and how engagement impacts student outcomes. Student engagement is an exciting area for educational researchers because it is a variable that schools have the ability to measure and potentially change, and research has shown that lack of student engagement is a good predictor of dropping out (Betts et al., 2010; Finn, 1993). Appleton et al. (2006) conceptualized student engagement into four categories: academic, 2 behavioral, psychological, and cognitive. Another area with promising evidence is the use of Early Warning Systems (EWS) to gather up-to-date data on student risk factors, such as attendance, grades, and behavior referrals. The EWS framework helps to identify students who may be at risk for dropout and connect them with intervention programs and services. The current study examined student engagement and participation in school-based intervention and prevention programs of students in grades 9-12 in a large suburban high school using an EWS framework. The study's main research questions were designed to determine whether the school's prevention and at-risk programs positively impacted cognitive and psychological engagement, which in turn would have an impact on the Early Warning Indicators (EWIs) of attendance, grades, and behavioral referrals. The study's main focus was to examine how school-based at-risk programs potentially act as a protective factor for student engagement. This is important because few studies have assessed the impact of at-risk programs on student engagement and EWIs while also evaluating the effectiveness of the EWS. There were also secondary research aims. The participating high school implemented a universal transition program for the incoming 9th grade class at the start of the school year. The transition program, called Link Crew, used assigned peer mentors, who were 11th- and 12th-grade peers, to help provide support and insight to incoming 9th grade students. Since this is a newly implemented program, the previous year's 9th grade class did not have assigned peer mentors. To assess the impact of the transition program, the study compared EWIs and student engagement between the 9th-grade students and 10th-grade students (the previous year's 9th-graders). The study also assessed for any 3 improvement in EWIs and student engagement variables from January to May for both groups of students (9th- and 10th-graders). Since previous research has found that there are differences in dropout rates based on income, race/ethnicity, and other factors, the study also examined the influence of these demographic variables. United States' Dropout Statistics and Trends The federal law, No Child Left Behind Act (NCLB), instated in 2002, placed an increased importance on school accountability. This necessitated that schools pay closer attention to their graduation numbers, and make steps towards increasing the academic performance of their students (Swanson, 2004). This also paved the way for discussions on a number of issues, such as, how to measure academic achievement and ability; differentiating achievement and ability levels for more disadvantaged subgroups; how to measure graduation rate accountability; and how to best measure graduation rates (Swanson, 2004). There are a number of ways to analyze graduation and dropout data. It is important to understand the difference between methods and how graduation and dropout rates are measured. One way the National Center for Educational Statistics (NCES) analyzes data is to use the averaged freshman graduate rate (AFGR), which takes the number of graduates in any given year divided by the estimated freshman enrollment 4 years previous (Aud et al., 2013; Chapman et al., 2011). There is also a similar method used to account for individuals who may transfer in or out of the school sample, called the adjusted cohort graduation rate (ACGR) (Stetser & Stillwell, 2014). Another estimate of graduation rate uses event dropout rate, which estimates the percentage of high school students who left school from the beginning of one school year to the beginning of the 4 next year without earning a diploma or alternative credential (Chapman et al., 2011). Status dropout rate is also used, which is the percentage of individuals within an age range who are not enrolled in school, and have not earned a high school diploma or alternative degree. Data can also be analyzed from the perspective of high school completion rates rather than dropout rates. Information regarding graduation and dropout rates is gathered using a number of data sources including, the Current Population Survey (CPS), the Common Core of Data (CCD), and the American Community Survey (ACS). These data are centralized on the EDFacts Collection System to help streamline data analysis at the state and district level in order to provide assistance and shape policy (Stetser & Stillwell, 2014). The different methods for analyzing graduation and dropout rates use a variety of information from the different databases (Chapman et al., 2011). It is important to understand the different graduation and dropout data collection and analysis methods in order to help clarify the statistics discussed throughout this paper. For example, using AFGR to measure graduation rates, Aud and colleagues (2013) report that the graduation rate in the U.S. has steadily increased from an estimated 73.7% of students in the 1990-91 school year to 78.2% in the 2009-10 school year. To put that in perspective, 3.1 million public high school students graduated on time with a regular high school diploma during the 2009-10 school year (Aud et al., 2013). This number remained relatively stable in the 2010-11 school year with an ACGR of 79.0% and an ACGR of 80.0% in the 2011-2012 school year (Stetser & Stillwell, 2014). This increase in graduation rates occurred across races and socio-economic statuses (SES). The gap between graduation rates of students from high-income and low-income families has significantly decreased from 1970 to 2011. The largest narrowing of the gap occurred 5 within the last 20 years from a 21-percentage point difference in 1990 to an 11- percentage point difference in 2011 (Aud et al., 2013). In 2009, there were an estimated 3 million status dropouts in the United States, which represents about 8.1% of the 38 million 16- through 24-year-olds living in the United States that year. It is important to note that much like previously stated dropout statistics, status dropouts have decreased from 14.6% in 1972 to 8.1% in 2009 (Chapman et al., 2011). The number of status dropouts decreased even further in 2012 to only 7% indicating continued improvement (U.S. Department of Education, National Center for Education Statistics, 2014). This downward trend is fairly consistent across race/ethnicities; however, for the Hispanic sample this decline did not occur until the 1990s (Chapman et al., 2011; U.S. Department of Education, National Center for Education Statistics, 2014). The status completion rate, which represents the number of 18- to 24-year-olds who are no longer enrolled in school, but do have a high school diploma or equivalent, also showed a positive upward trend of individuals earing a high school credential of 89.8% in 2009 (Chapman et al., 2011). The graduation completion numbers have increased in part because of alternative education options. The alternative education movement gained momentum in the 1970s, and has helped to provide alternate pathways for students considered at-risk or in need of a different education model than the general student population (Stanley & Plucker, 2008). There are also alternative credential pathways that students can take if they have not completed high school, such as the General Education Development (GED). The GED has been around since 1942 and has continued to evolve with the secondary education curriculum. The GED test is highly accessible. For example, in 2012, 702,000 6 adults or 1.5% of individuals without a high school diploma took at least one GED test and 74.6% of individuals who took the exam for the first time passed (GED Testing Service, 2013). Although the GED does improve outcomes for high school noncompleters, it is important to note that individuals with a GED certificate often still do not fare as well as individuals with a traditional high school diploma (Chapman et al., 2011). The NCES also reports graduation rates by race/ethnicity. For the 2011-12 school year, graduation rates using AFGR by race/ethnicity were as follows: Asian/Pacific Islanders (88.0%), Whites (86.0%), Hispanics (73.0%), American Indians/Alaska Natives (67.0%), and Blacks (69.0%) (Stetser & Stillwell, 2014). With the national average graduation rate around 80.0%, it is clear that minority status has an effect on graduation rate since the graduation rates of Hispanics, American Indians/Alaska Natives, and Blacks all fall below the national average. There are also statistical differences between genders. According to the 2011-12 data, female students' graduation rates are at 85.0%, which is about 7 percentage points higher than male students at 78.0% (Stetser & Stillwell, 2014). These data suggest that the public education system has not been entirely successful in supporting all students to graduate; however, there have been strides in decreasing differences in the graduation rate between students from different races and socio-economic statuses. Event dropout rates have also trended downward since 1972 (6.1%); however, there was a brief spike between 1990 and 1995 when the rate began to increase again before trending back down to 3.4% from October 2008 to October 2009 (Chapman et al., 2011). There was no reported difference in dropout rates by gender from 1972 to 2009; 7 however, the years 1974, 1976, 1978, and 2000 all had higher event dropout rates among males. The event dropout rate was higher for Black (4.8%) and Hispanic (5.8%) students than White (2.4%) students in 2009. There were some interesting differences between races/ethnicities in regard to event dropout. Black students experienced a decline from 1972 to 1990, then increased from 1990 to 1995; however, after 1995 the rates have fluctuated with no measurable trend. Hispanic students had no measurable trend from 1972 to 1995, but event dropout rates declined from 1995 to 2009. Among White students, event dropout rates mirrored the overall population trends that were previously stated (Chapman et al., 2011). Graduation and status dropout rates are also reported by state, and vary considerably. The status dropout rate across the U.S. averaged around 3.3% in the 2010- 11 and 2011-12 school years (Stetser & Stillwell, 2014). The status dropout rates by state reveal a wide range with Alaska having the highest status dropout rate at 6.9% and New Hampshire having the lowest at 1.3%. Different conclusions can be made when examining the ACGR data from the 2011-12 school year by state, which suggests that the District of Columbia has the lowest graduation rate at 59.0%, and Iowa has one of the highest graduation rates at 89.0% (Stetser & Stillwell, 2014). The current study took place in Utah, which had a 76.0% graduation rate using ACGR in 2011-12 and a status dropout rate of 1.5% (Stetser & Stillwell, 2014). The data suggest that Utah is below the national average for graduation rates; however, Utah's status dropout rate is lower than the national average. Overall, the literature suggests that graduation rates continue to increase while dropout rates are decreasing. This is a step in the right direction, but there are still a large 8 number of students who fall short of earning their diplomas every year. As previously stated, there are differences in graduation rates between race/ethnicities and income levels, but there are two other groups that consistently fall well below the national average and other disadvantaged groups. They are English Language Learners (ELL) and students with disabilities, with completion rates, using ACGR, at 59.0% and 61.0% respectively (Stetser & Stillwell, 2014). Schools need to identify ways to continue to increase their graduation rates, especially among minority and disadvantaged groups since a substantial gap in graduation rates persists between different races/ethnicities, income brackets, and other disadvantaged groups. All students should be able to obtain a high school diploma because without a high school degree or equivalent these individuals will be at a significant disadvantage, and at a higher risk for poorer life outcomes. Negative Impact of Dropping Out The issue of high school dropout has been referred to as a national crisis. Students who fail to complete high school are missing out on a major milestone in their educational careers, and subsequently fail to gain the economic and social advancement benefits that a high school diploma has to offer. The students who drop out of high school will be at a disadvantage, not only because they did not earn a diploma, but they are also less prepared for their future. It has been shown that these students may not work as many hours or earn as much money as high school graduates, which is only exacerbated by the growing economic trend to hire more educated workers. Individuals who have dropped out of high school are more likely to be on public assistance, use drugs, be arrested, and possibly spend time in prison (Dynarski & Gleason, 2002). To be more specific, the National Dropout Prevention Center/Network (2009) reports that high school 9 dropouts in the United States earn about $9,245 less every year than individuals who complete high school, and the unemployment rates are almost 13 points higher for high school dropouts. Furthermore, their average salary per year throughout their life is much lower than the average for high school graduates. In 2009, the average income for individuals aged 18 to 67 who did not complete high school was estimated to be around $25,000 (Chapman et al., 2011). To put the salary estimate in perspective, the salary of an individual who completed a GED certificate was estimated at $43,000, which is close to double that amount (Chapman et al., 2011). The significant wage gap for individuals without a high school education has been steady for more than 10 years (Aud et al., 2013). As previously stated, unemployment is a significant problem for high school dropouts. The job market is particularly tough for individuals with less than a high school diploma, who often struggle to find employment compared to those who have a high school diploma and higher. In 2012, the employment rate for young adults aged 20-24 with a high school diploma was 64% compared to 48% for young adults without a high school diploma (Aud et al., 2013). The employment rate did not increase much for older individuals without a high school diploma: 56% for adults aged 25-34, and 53% for adults aged 25-64 (Aud et al., 2013). Of note, the employment rate was significantly lower for women with less than a high school completion across all of the age groups (Aud et al., 2013). Aside from struggles to find employment, individuals who do not graduate from high school tend to be in poorer health than individuals who completed high school, regardless of income (Pleis, Ward, & Lucas, 2010). The health issues of high school 10 dropouts include higher rates of different types of reported body pains, migraines, diabetes, ulcers, kidney disease, liver disease, hearing trouble, vision trouble, absence of all natural teeth, and reported symptoms of anxiety and depression (Pleis et al., 2010). Furthermore, those who do not graduate from high school report more trouble with different types of physical activities, and have the lowest percentage of engagement in vigorous physical activity (Pleis et al., 2010). Also, considering their difficulties with employment obtainment and stability, these individuals are the least likely to have a regular place for medical care (Pleis et al., 2010). High school dropouts are also at risk for a number of other factors that could significantly impact their quality of life. For example, individuals who have dropped out of high school are more likely to become teen parents (National Dropout Prevention Center/Network, 2009). Unfortunately, they are also at a much higher risk of being institutionalized or imprisoned. In 2009-2010, it is estimated that approximately 40% of 16- to 24-year-olds in the institutionalized population (including prison inmates) were high school dropouts (Aud et al., 2011). Individuals who drop out of high school can also affect the United States' economic advancement. Education is an important factor in creating a well-educated workforce that can advance a nation's economy. This is especially critical with increased competition in global market places. Riggs, Carruthers, and Thorstensen (2002) estimated that high school dropouts potentially cost the United States $24 billion each year because of crime involvement, food stamps, housing assistance programs, and Temporary Assistance for Needy Families (TANF) benefits (as cited in Porowski et al., 2011). Another group of researchers, the Alliance for Excellent Education (2011), looked at the 11 potential economic benefits of decreasing dropout rates. These researchers hypothesize that cutting the dropout rate of a single high school by half would support nearly 54,000 new jobs and could increase gross domestic product (GDP) by as much as $9.6 billion. It is clear that high school dropout in the United States is a source of major financial loss. With the financial cost of dropout being so high, it is no wonder research in the field of dropout prevention and intervention is booming. The cost benefits from implementing effective programs are a major driving force in the research. Theories Behind High School Dropout It is clear that the impact of high school dropouts has negative outcomes for not only the individuals who drop out, but also for society as a whole. For this reason it is important to try to determine why students drop out and how to prevent dropping out; however, there are a number of different factors both proximal and distal that affect dropout (Rumberger, 2001). Due to the complexity of why students drop out, many researchers have posited theories to better explain and predict dropping out by looking at common themes and variables. There are a number of theories and models for student dropout. For example, Finn (1989) proposed two different theories for why students drop out of high school: the Frustration-Self-Esteem and the Participation-Identification models. The Frustration-Self- Esteem model states that academic failure is the tipping point in a student's academic career that causes the student to either reject school and/or be rejected by the school. The student then internalizes the feelings of frustration and embarrassment from the school failure, and this creates an impaired self-view. The student views the school as being responsible for his/her failures by not providing engaging instruction and/or a healthy 12 learning environment in which the student felt emotionally supported. The student then acts out toward the school, which could take the form of skipping school, disrupting class, or other acts that could lead to a school discipline referral. Finn's (1989) other theory, the Participation-Identification model, focuses on how students' school involvement or participation can impact their behavioral and emotional well-being, and in-turn impact school outcomes. Finn's (1989) model views school participation as an ongoing process. This model hypothesizes that participation in school-based activities increases the likelihood that a student will identify with the school and enjoy school. Therefore, the more they participate and enjoy school, the better outcomes for students and the more likely they will complete school. The opposite would then be true if a student was a nonparticipator; that student would then have poorer school performance and disengage from school, and the likelihood of that student not completing school would increase. Finn (1989) further goes on to state, "the ability to manipulate modes of participation poses promising avenues for further research as well as for intervention efforts." (p. 117). The idea that participation can be manipulated offers an avenue for school professionals and researchers to intervene and help students to become more involved and connected with school (Finn, 1989). The Participation- Identification model also posits that there are student factors that can impact student success such as skills, attitude, and behavior that are formulated prior to formal schooling. This will be discussed more in the Risk Factors for Dropout section. Finn's (1989) theories, as well as those of other researchers (Newmann, 1992), set the stage for current research in the area of student engagement. For example, Rumberger (2001) identifies two additional frameworks for high school dropout: an individual 13 perspective and an institutional perspective. The individual perspective focuses on a student's values, attitudes, and behaviors, and how those attributes impact a student's decision to leave school. Again, this framework has an emphasis on student engagement and how over time, different variables affect a student's engagement in school, which then impacts the student's perspective and beliefs about school. The institutional perspective takes into consideration a shift in the field of psychology away from focusing solely on the individual and thinking about individuals in terms of the contexts in which they live (Rumberger, 2001). This perspective considers family, school, community, peers, and demographic factors. Rumberger (2001) argues that a student's environment, namely family and community, has a strong influence on whether a student will drop out of school, but that many of the negative effects of a student's environment outside of school can be mediated by school factors. There are four types of school characteristics that have been shown to impact student performance: available resources, student composition, structural characteristics of the school, and a school's processes and practices (Rumberger, 2001). Of the four school characteristics, a school's processes and practices is the one characteristic that schools can easily modify. A school can aim to improve its overall effectiveness by shaping policies that help to keep students engaged and successful. A common theme across these theories and frameworks is the importance of student engagement in preventing dropout, and the importance of considering not only student factors but also environmental factors. Student engagement theory is currently considered the primary model for understanding student dropout (Christenson et al., 2012). Furthermore, the early research on dropout prevention has led many researchers to 14 now focus more on a school completion perspective. From this perspective, interventions can be directed toward helping students gain the skills necessary to be successful in school and have the skills needed to prepare them for postsecondary success (Reschly & Christenson, 2012). Student Engagement Since Finn (1989) prosed the Frustration-Self-Esteem and the Participation- Identification models, there has been an increased research interest in student engagement due to (1) a need to better understand engagement in the scope of dropout prevention and intervention, (2) attempts to incorporate engagement into a general school reform model, and (3) expanding research on the role of motivation (Reschly & Christenson, 2012). However, research in all of these areas becomes complicated by differences in definitions of student engagement. One of the reasons for the differences in opinion is that educational and motivational research is studied within different academic fields, including educational psychology, which is more applied in focus, and developmental psychology, which is more theoretical. For example, there are researchers who view engagement along a continuum from high engagement to disengagement, and there are others who believe disengagement or disaffection should be treated as separate continua (Reschly & Christenson, 2012). There is also a debate on whether it is important to separate facilitator (context) and indicator (student) variables when studying student engagement (Reschly & Christenson, 2012). Due to these different perspectives, the definition of student engagement may be viewed as muddled and what exactly makes up the construct of student engagement remains in contention. It is clear, however, that student engagement is an important area for continued 15 research. The reason researchers and policy makers continue to study student engagement is summed up nicely by Newman (1992), who states, "The most immediate and persisting issue for students and teachers is not low achievement, but student disengagement." (p. 2). This statement reiterates that student disengagement is a major concern to be addressed through intervention. As previously stated, however, student engagement has numerous definitions, which makes studying the construct challenging. What researchers agree on is that student engagement encompasses multiple factors and it is a complex phenomenon. For example, Balfanz, Herzog, and Mac Iver (2007) define student engagement as, "a higher order factor composed of correlated subfactors measuring different aspects of the process of detaching from school, disconnecting from its norms and expectations, reducing effort and involvement at school, and withdrawing from a commitment to school and to school completion." (p. 224). This definition reveals the complexities of student engagement, but also speaks to the importance of a student being engaged and being connected and committed to school. To define student engagement further, engagement represents a student's active involvement in school tasks or activities. Reeve and colleagues (2004) also state that engagement refers to a person's behavioral intensity and emotional quality while he or she is participating in a task. The emotional quality aspect of engagement is important to consider. When students or any individuals experience positive emotions it is believed that this helps them to expand their thoughts and behaviors more adaptively to their environments and cultivate continued well-being (Fredrickson, 1998, 2001). When individuals experience negative emotions it is believed to have the reverse effect, and decrease learning and adaptive thoughts and behaviors (Fredrickson, 2001). Therefore, positive emotions can lead to 16 increased coping skills and are correlated with success across contexts, including school (Fredrickson, 2001; Reschly et al., 2008). Student engagement can also be considered the binding agent of multiple contexts, including home, school, peers, and community (Reschly & Christensen, 2012). Fredricks and colleagues (2004) conceptualize engagement more as a metaconstruct that connects multiple areas of research and also subsumes motivation within the construct of engagement. Even though student engagement is multifaceted and complex, and the definition may not always be agreed upon, research is needed to continue to explore the avenues in which engagement can be measured and used to predict academic outcomes (Reeve et al., 2004). As previously stated, there are many researchers with the perspective that student engagement is an important area of research in dropout prevention and intervention research. Finn (1989) first advocated for student engagement to be considered in dropout prevention research, and identified it as an important construct that needed further exploration. Many years after Finn (1989), Appleton, Christenson, Kim, and Reschly (2006) demonstrated that engagement could be altered through interventions. Another group of researchers, Fredricks, Blumenfeld, and Paris (2004), agreed with Appleton and Colleagues (2006) and provided evidence that engagement is malleable and responsive to contextual and environmental change. With this mindset, student engagement is a potential predictor of dropout, and school personnel can effectively intervene when student engagement is low. With continued research and support, student engagement continues to be at the forefront of research and reform in the field of dropout prevention. As first stated by Finn (1989), engagement can be divided into two subtypes: behavioral and affective. Behavioral engagement is described as participating in class and 17 school, and affective engagement is described as identifying with school, valuing learning, and belongingness. The literature continues to expand and so has the theoretical model of engagement. Recent reviews of literature propose that engagement has three subtypes; a third cognitive subtype has been added to the original behavioral and affective subtypes. The cognitive engagement subtype includes self-regulation, setting goals for learning, and being invested in one's education (Fredricks et al., 2004). Most recently, Appleton et al. (2006), prominent researchers in the field of student engagement, have broken down engagement into a measurable multifaceted taxonomy encompassing academic, psychological, behavioral, and cognitive components. They based their model on previous theoretical engagement models and 13 years of intervention research in the schools using Check and Connect (Christenson, Stout, & Pohl, 2012). The Appleton et al. (2006) model is set up not only to measure student engagement based on the four factors, but also to better understand goodness-of-fit between the student, the student's environment, and factors that may impact learning. Academic, behavioral, psychological, and cognitive factors include a number of indicators for each engagement subtype (Appleton et al., 2006). For example, academic engagement is comprised of the variables: homework completion, on-task rate, and credits toward graduation. In the Appleton et al. (2006) model, behavioral engagement encompasses the measureable variables: attendance, suspensions, classroom participation, and involvement in extracurricular activities. Cognitive engagement indicators are more internal in nature, and include self-regulation, being able to connect schoolwork to future goals, valuing learning, setting personal goals, and autonomy. Psychological engagement is also internal in nature, and includes feelings of belongingness or identification with 18 school and peers, and relationships with teachers and peers. The four engagement factors (academic, behavior, cognitive, and psychological) in the Appleton et al. model (2006) can be affected by contexts such as school, community, peers, and/or family. For example, within the school context there are a number of variables that could impact engagement such as the school climate, mental health supports, instructional programs, learning activities, clear and appropriate classroom expectations, structure, and student-teacher relationships. Peers also have various ways of impacting a student's engagement. Peers can influence educational expectations, common interests and values about school, attendance, beliefs about academics, effort, and peer aspirations for learning (Appleton et al., 2006). Furthermore, family variables such as goals and expectations for one's children, supervision, educational resources in the home setting, and support for learning, have a major impact on engagement. The variables described within each context could be viewed as alterable predictors of dropout, and may act as protective factors for academic success and school completion (Reschly & Christenson, 2006). The multivariable taxonomy of contexts that influence the specific engagement domains sets up a useful framework to identify potential risk factors and areas for intervention (Appleton et al., 2006). All of the domains of engagement are important to consider, but academic and behavioral engagement are the easiest to observe and measure (Appleton et al., 2006). Although difficult to measure, research continues to emerge that shows support for the cognitive and psychological domains of engagement. For example, a study done by Greene, Miller, Crowson, Duke, and Akey (2004) showed that students' perceptions about the classroom structure were important factors in motivation. Another important 19 finding was that students' perceptions of classwork being important for their future goals and success also affected their cognitive engagement. Greene and Miller (1996) also found that college students' perceived ability and learning goals correlated with the use of meaningful cognitive engagement as measured by reports of high levels of self-regulatory activities and higher use of meaningful study strategies. The study also found that higher levels of cognitive engagement increased academic achievement (Greene & Miller, 1996). There are limitations in previous research in how cognitive and psychological engagement has been measured. For example, the same items on a measurement of engagement have been used to measure different subtypes of engagement, and subtypes of engagement have been analyzed separately without comparison to other subtypes. Furthermore, raters vary across studies, and studies that use observer report such as a teacher, are thought to be highly subjective (Appleton et al., 2006). The method of using informants other than the student is highly inferential and reports may vary by rater. In order to more reliably and validly measure the internal construct of engagement, student report is often preferred. The reason student self-report is considered more valid is that students have a better understanding of how their own contexts impact their experiences. Reschly and Christenson (2012) further suggest that students are able to accurately report on their own engagement and environments, and their perspective should be considered when choosing and implementing interventions. In order to more accurately capture cognitive and psychological/affective engagement, Appleton and Christenson (2004) created a valid self-report measure called the Student Engagement Instrument (SEI). The SEI treats student engagement as an 20 outcome variable, which it is, but it is important to understand that student engagement is also a process and acts as a mediator variable for academic and behavior outcomes (Reschly & Christenson, 2006, 2012). The current study uses the SEI to measure student engagement; however, it is important to note that there are numerous student engagement measures. Fredricks and colleagues (2011) conducted a review of 21 student engagement instruments. These instruments differ in the source of the data (student, teacher, observation), what subtypes of engagement are measured, and whether general or more specific forms of engagement are assessed (Fredricks et al. 2011). Advances in measurement and construct clarification would help to focus and expand student engagement research. There are still many areas of student engagement research that need to be expanded upon and clarified. For example, there is limited knowledge about the effectiveness of intervention outcomes. Engagement interventions include helping students to function better in their environments, changing curriculum and school structures, personalizing students' learning environments, and building relationships among students as well as between students and staff. With the variations in the focus of engagement interventions, it is important for future research to determine which interventions are the most effective, for which groups of students, and under what conditions. Furthermore, additional information about the optimal duration and intensity of interventions is needed. As the field of student engagement continues to expand, it will be important for researchers to work toward answering these questions (Reschly & Christenson, 2012). 21 Risk Factors for Dropout There are a number of risk factors that potentially impact student disengagement and a student's ability to be successful in school. Gathering information about potential risk factors is important because it could help identify areas where a school or other community agency may be able to intervene. To gain a better understanding of the research on dropout risk factors, Hammond et al. (2007) completed an extensive literature review spanning the years 1974-2002. To be included in the analysis, the studies had to directly analyze the data source, examine school dropout and/or high school graduation as the dependent variable, collect longitudinal data over a period of at least 2 years, examine a variety of predictor variables in different domains (individual, family, school, and/or community), use a multivariate statistical technique or model, and include a sample of 30 or more students classified as noncompleters. Of the 44 studies found, only 21 met the criteria to be included in the authors' analysis of at-risk variables. These studies used national data samples, community samples, and school district samples, and spanned different periods in time and diverse communities both in location (urban, rural, and suburban) and demographics (race/ethnicity, SES, and gender). In considering the data from the almost 30-year time period, Hammond et al. (2007) found 25 significant risk factors across eight different categories. The authors estimated that about 60% of the factors were considered individual factors and 40% were family factors. In addition to considering individual and family factors, similar to previous researchers, the authors further conceptualized these risk factors into four different domains: individual, family, school, and community (Rumberger, 2001). These domains encompass the many risk factors that can impact 22 students' school success, and provide researchers and practitioners with targeted areas in which to intervene. Risk factors rarely occur in isolation; therefore, it is important to consider all possible risk factors that could be impacting a student. The review found that there was not one single variable that could accurately predict dropout, but that prediction strength increased when multiple variables were considered (Hammond et al., 2007). Furthermore, the review aimed to better understand when each individual risk factor began impacting a student during their academic career (e.g., elementary school, middle school, and/or high school). Finally, the review stated that these risk factors do not usually have an immediate impact on student engagement, but rather a cumulative effect that builds over time (Hammond et al., 2007). Hammond et al. (2007) found that the studies on prevention and dropout were lacking in the rigor of their evaluation of program effectiveness, and the studies collected little to no long-term data. This significantly limited the number of studies that met criteria for the U.S. Department of Education's What Works Clearinghouse, a group designed to review and compile data on best practices in dropout prevention and intervention. From the results of the review, Hammond et al. (2007) concluded that there are a number of individual factors that put students at risk for dropout. First, a student's family background is a powerful predictor of student dropout. A student's family background has a profound impact on who students are as individuals. It is important to think about students in terms of their family background because previous studies have found that family factors were the single most important factor in school success; however, subsequent research has found that schools can act as protective factors to mediate family 23 background as a risk factor (Rumberger, 2001). The individual/family characteristics that place a student at greater risk are low socioeconomic status; low parental education level, and/or occupation; and family structure (Rumberger, 2001). Hammond et al. (2007) also reported that a large number of siblings, frequent mobility, family conflict, and not living with both birth parents were potential risk factors for students. Many of these risk factors do not occur in isolation. For example, Jerald (2006) stated that research has consistently shown that minority students and students living in poverty are more likely to drop out of school. Moreover, minority students may also be English Language Learners (ELL); having limited English proficiency is another potential risk factor for dropout (Jerald, 2006; Rumberger, 2001). Immigration status further places a student at risk for dropping out (Rumberger, 2001). ELLs have a number of challenges that potentially impact their education. First and foremost, ELLs need to learn to speak English, and write and read in English, and master English well enough to participate in different academic subjects needed for graduation (Gwynne, Stitziel Pareja, Ehrlich, & Allensworth, 2012). Another challenge for older ELL students is that until recently many school-based intervention programs for ELL students occurred primarily in elementary school settings, although there is now a push to meet the growing needs of ELL students in the middle and high school grades (Capps et. al., 2005). ELL students typically have lower academic grades, and earn fewer credits in their core classes than their non-ELL peers, placing them at an even greater risk of dropping out than other minority students (Chapman et al., 2011). As previously stated, students from low-income families are also at an increased 24 risk for dropout. They are about two times more likely to drop out, at a dropout rate of 10%, than middle class families with a dropout rate of 5.2%, and almost 10 times more likely to drop out than students from high income families with a dropout rate of 1.6% (U.S. Department of Education, 2014). Families with high financial burdens may also be more likely to have a low commitment to education. For example, families may set low academic expectations for their child, have a sibling or siblings that have dropped out, show little engagement with the school, and have little to no conversations about school in the home (Hammond et al., 2007). Another individual factor that also places a student at higher risk for not completing school is having a learning disability (Deshler et al., 2001). Students with learning disabilities struggle to achieve academically and these struggles continue to worsen as the content in classrooms becomes more academically complex, and schools struggle to accommodate and meet these students' individualized needs (Deshler et al., 2001). Furthermore, students with learning disabilities also tend to have an increased number of behavior problems compared to students without disabilities (Sabornie & deBettencourt, 2004). This combination of behavioral and academic problems places these students at a greater risk for not completing high school. Reschly and Christenson (2006) also examined student engagement differences, a well-researched variable that can impact high school dropout, of students with learning disabilities and emotional or behavior disorders. They found that individuals with learning disabilities and emotional or behavior disorders scored much lower on engagement measures than their average-achieving peers, which in turn increases their risk for dropping out. Additionally, individuals with a disability are more likely to have a combination of risk factors 25 compared to other students (Wagner et al., 2006). Another individual risk factor that has a significant impact on dropout is taking on more adult responsibilities, such as having a high number of work hours or having a child while still attending school (Hammond et al., 2007). Although males are at a higher risk of dropout than females, teen pregnancy is a very strong risk factor for females, especially in the United States (Wilson et al., 2011). Perper, Peterson, and Manlove (2010) report that only around 50% of teen mothers in the United States receive a high school diploma by the age of 22. A student's peer group also has a great influence on the student's potential risk for dropping out of high school. If one's peer group is high-risk, such as a gang and/or low achieving, a student could be engaging in high-risk behaviors that could place him/her at a higher risk for dropout (Hammond et al., 2007). Peer groups could also lead to having a poor attitude toward school. Students with early antisocial behaviors including violence, substance abuse, other criminal offenses, as well as early sexual experiences, are more at-risk for dropout (Hammond et al., 2007). Poor academic performance is one of the most consistent variables that places a student at risk for dropout. It is one factor that has been found to have an early impact, as early as first grade (Alexander, Entwisle, & Kabbani, 2001), and one of the most frequently reported reasons why students leave school (Hammond et al., 2007). Grade retention, which is related to poor school performance, has been repeatedly shown to result in poor outcomes for students (Alexander et al., 2001; Rumberger, 2001). Retention at any grade can be detrimental, and grade retention has been shown to have an additive effect, in that the greater number of times a student is retained, the poorer the 26 outcomes (Alexander et al., 2001; Gleason & Dynarski, 2002). Lack of school engagement can be manifested as poor attendance, lack of effort, no involvement in extracurricular activities, lack of commitment to school, and low academic expectations, all of which are significant risk factors. This is clearly demonstrated in the study completed by Finn and Rock (1997), which examined moderating and mediating variables of at-risk groups to see why some students were more academically successful within these groups. They found that after controlling for a number of variables (socioeconomic status, parent school, family structure, etc.) that teacher- and self-reported engagement were significantly correlated with better outcomes. This is an important finding because student engagement is a variable that has the potential to be alterable within the school environment, unlike a demographic variable such as socioeconomic status. For these reasons, monitoring student engagement provides a promising approach to intervention and prevention of dropout (Reschly & Christenson, 2006). As previously noted, dropout should be viewed as a gradual process of many events and factors cumulating that can lead to a student withdrawing from school. Reschly and Christenson (2006) purport that this gradual dropout process can best be explained using theories of student engagement. A majority of risk factors have a significant impact on dropout in middle and high school; however, student performance variables, poor attendance, school behavior, and family background characteristics were found to significantly impact dropout as early as elementary school (Hammond et al., 2007). For example, a primary risk factor is school misconduct, which was found to be even more predictive of dropout when aggression and misconduct occurred at an early 27 age (Hammond et al., 2007). Another important time period when risk factors can significantly impact student outcomes, is the transition from middle school to high school. The 9th-grade year often sets the tone for a student's academic success, as well as success beyond high school. The National High School Center (2012, October) reports that more students fail 9th grade than any other grade, suggesting that it is very difficult for students to recover from a failed 9th-grade year. The large number of risk factors that could potentially place a student at risk can make intervention efforts challenging. For this reason, it is important to know and understand the most prominent risk factors and how to identify them. The ability to monitor and track risk factors, including student engagement, can facilitate improved intervention efforts. Early Warning Systems With a multitude of risk factors having such a strong influence on dropout, it is important for school staff to be knowledgeable about potential risk factors and have a system in place to identify risk factors. An Early Warning System (EWS) helps school staff to regularly access student data to accurately identify risk factors, and target those students in need of specialized supports. This is made easier with new technological advances. Schools now have the capability to track and analyze student data through computer software and subsequently, can more easily identify students who may be at-risk for school dropout and in need of intervention or supports. An EWS uses technology to identify and track dropout-specific variables such as poor grades, low attendance, and behavior incidents (Davis et al., 2013; Frazelle & Nagel, 2015). The risk factors that are tracked in an EWS are referred to as Early Warning Indicators (EWIs) (Davis et al., 28 2013). The main purpose of the EWS is to use student data to identify students who may be off-track for graduation, and provide supports and interventions to help students back on-track. The EWS can also help identify on-track patterns among students (Frazelle & Nagel, 2015). In theory, an EWS is a more collaborative approach in which administrators, teachers, and parents or problem-solving teams can use data to assess whether students are on track or in need of additional resources and/or interventions (Neild, Balfanz, & Herzog, 2007). In practice, however, it is important for an EWS to be well organized and efficient. Frazelle and Nagel (2015) identified five main components that are needed to implement a successful EWS: (1) creating and training a team to use the EWS, (2) identifying accurate EWIs, (3) designing and using reports, (4) using appropriate interventions for each individual student, and (5) evaluating progress and effectiveness of interventions. Kennelly and Monrad (2007) specified detailed steps to consider when setting up an effective EWS. Their suggestions are based on considerable research and aim to identify students even earlier than high school age, when possible: 1. Establish a data system that tracks individual student attendance, grades, promotion, status, and engagement indicators, such as behavioral marks, as early as fourth grade. 2. Determine criteria for who is considered off-track for graduation and establish a continuum of appropriate interventions. 3. Track ninth grade students who miss 10 days or more of school in the first 30 days (Neild & Balfanz, 2006). Even moderate levels of absences are a cause for concern. Just one to two weeks of absence per semester - which was typical for freshmen participating in a key Chicago study - was found to be associated with a substantially reduced probability of graduating (Allensworth & Easton, 2007). 4. Monitor first quarter freshman grades, paying particular attention to failures in core academic subjects. Receiving more than one F in core academic subjects in ninth grade- together with failing to be promoted to tenth grade - is 85% 29 successful in determining who will not graduate on time (Allensworth & Easton, 2005). Schools can offer immediate academic supports to the students who are failing in the first quarter of freshman year. 5. Monitor fall semester freshman grades, paying particular attention to failures in core academic subjects. As first semester grades are posted, schools can develop individual student dropout strategies. By the end of the first semester, course grades and failure rates are slightly better predictors of graduation than attendance because they indicate whether students are making progress in their courses (Allensworth & Easton, 2007). 6. Monitor end-of-the-year grades. The end-of-the-year grades will provide further information about failure rates and reveal grade point averages, providing detailed information about who is likely to struggle in later years and is considered by some researchers to be the best indicator for predicting nongraduates (Allensworth & Easton, 2007). In general, grades tend to be a more accurate predictor of dropout than test scores. Track students who have failed too many core subjects to be promoted to tenth grade. This provides perhaps the most critical information about which students should receive specialized attention and support. Research has shown that those who fail to be promoted are more likely to drop out. According to Alexander, Entwistle, and Horsey (1997), being held back trumps all for dropout indicators. (pp. 7-8) Kennelly and Monrad (2007) mention several EWIs to track. Kennelly and Monrad's (2007) steps also suggest that EWIs can be grade specific. For example, Balfanz et al. (2007) found that around 60% of future dropouts can be identified with accuracy as early as 6th grade, using only four major indicators: poor attendance, poor behavior marks, failing math, or failing English. Interestingly, the study also confirmed that students who drop out do so in different but identifiable ways. They found that in their 6th-grade sample, the most common occurrence was students having the risk factors of either poor behavior or low attendance; or two factors, especially poor behavior combined with failing one of their core classes (Balfanz et al., 2007). Another study found that 22% of 9th-grade students who did not have enough credits to be promoted to 10th grade went on to graduate from high school in 4 years (Allensworth & Easton, 2005, 2007). The same study also found that on-track status was a more powerful predictor of 30 high school graduation than test scores and demographic characteristics combined (Allensworth & Easton, 2005, 2007). The results from this study confirm the importance of the 1st year in high school. The 9th-grade transition year can set the stage for whether or not a student drops out of high school or does not complete high school in the traditional 4 years. Previous EWS research should act as a guide; however, a school community should identify the specific EWIs that are the most accurate indicators for their students. Frazelle and Nagel (2005) suggest that school personnel examine student data from previous school years to identify the EWIs that are most related to student performance and graduation. The University of Chicago Consortium on Chicago School Research (2014) also recommends that EWIs should be valid for the intended purpose, actionable, meaningful and easily understood, and match district and school priorities. The Consortium also suggested that EWS teams only use indicators that the school has control over. Tracking indicators such as family factors should be avoided although there is a strong correlation between these factors and school completion. Frazelle and Nagel (2015) suggest that EWS teams start with a base set of indicators such as attendance, behavior, and academic performance in classes. School teams can then add indicators that are unique and helpful to identify their own students who are off-track. Also, end of year assessment scores may not be the best EWIs; student progress can be better assessed in shorter, more measurable intervals (Frazelle & Nagel, 2015). An EWS needs to be implemented with fidelity, and requires support from administration and the school community. EWSs are being used across the country, and are being implemented in a number of different ways. For example, Sioux Falls District 31 uses teams at the district level to track students' progress toward graduation, while Houston Independent School District has EWS teams at the school level that incorporate other community resources (Frazelle & Nagel, 2015). The National High School Center recommends a mixed-level team approach with stakeholders at the district and school-level, and incorporating staff from not only the high schools, but also stakeholders from middle schools that feed into the high schools. Some other considerations when implementing an EWS are to make sure roles are clearly defined, and create S.M.A.R.T. (specific, measurable, achievable, relevant, and time bound) goals (Kekahio & Baker, 2013). Johns Hopkins University School of Education, Center for Social Organization of Schools (2010) has a specific breakdown of how an EWS meeting should be set up and run. Furthermore, school districts should aim to provide initial and continued professional development to help the EWS teams work more effectively. Also, involving community organizations can be useful to help ease the workload of school staff. The school district where the present study will take place uses problem-solving teams with teachers, administrators, and other school staff members to analyze data and come up with school-based solutions. There is also district level department that provides professional development and support to help school teams use data to make data-driven decisions. There is a great deal of information on how to set up an EWS; however, the data examining the effectiveness of EWSs are still somewhat limited (Frazelle & Nagel, 2015). Considering the long history of dropout research, it has taken some time to draw conclusions about how to track and intervene with students who are at-risk for not completing high school. Researchers, such as Jerald (2006), have called into question 32 why there has been a lack of research on EWIs, especially after over 40 years of documented concern over student dropout. Fortunately, the analysis of EWIs in conjunction with EWSs has significantly increased over the past 5 years (Davis et al., 2013) and likely will continue. Prevention and Intervention Programs Once students and specific risk factors have been identified using an EWS it is important to provide students with appropriate supports or services through intervention programs that are going to meet their unique needs. This can be difficult if a school does not have a number of different programs to meet all of the needs of their student body. Along with having programs tailored for the needs of at-risk students to prevent dropout, these programs need to be effective and administered with efficacy. This has been a criticism because school districts have used intervention and prevention programs for years, but there have not been many systematic studies on these programs' effectiveness. Gleason and Dynarski (2002) suggest that effective intervention and prevention programs may be few in number because it is difficult to match programs to students' unique psychosocial and academic needs. They go on to argue that many dropout prevention programs are one-size-fits-all, and are not effective for all students (Gleason & Dynarski, 2002). Schools should consider students' individual risk factors and try to match them with appropriate prevention services. If schools do not have access to a particular dropout prevention program they should investigate prevention and intervention strategies that have been proven to be effective with different at-risk populations (Gleason & Dynarski, 2002). Hammond et al. (2007) from the National Dropout Prevention Center/Network 33 and Communities In Schools, Inc., did an extensive literature review to identify effective intervention and prevention programs for a wide range of at-risk students. In order to meet criteria as exemplary, the program had to be ranked in the top tier or level by at least two sources, currently be in use, have no major revisions since the program was ranked, have consistently positive outcomes, and focus on school-aged children in grades K-12. The review identified 50 evidence-based programs with many programs tailored toward specific at-risk student populations. The review cautioned that there are a number of flaws in dropout prevention program research such as: there has been little rigorous evaluation of program effectiveness, there is a lack of longitudinal data, and few programs meet the criteria for the U.S. Department of Education's What Works Clearinghouse (Dynarski et al., 2008). The criteria for U.S. Department of Education's What Works Clearinghouse for a classification of having "Strong" evidence are programs and/or practices that have both high internal validity and external validity (Dynarski et al., 2008). This means that the study design needs to be able to support conclusions, and that results generalize to multiple settings and participants. The study designs that are recommended are well-designed randomized controlled trials or quasi-experimental (without randomized control). Further, evidence is stronger with multiple studies, and there should be no studies that contradict the results (Dynarski et al., 2008). The U.S. Department of Education's What Works Clearinghouse outlines six recommendations for dropout prevention (Dynarski et al., 2008). Their first recommendation is to use data systems or an EWS to estimate the number of students who are noncompleters and identify individual students who are at-risk. Next, they 34 suggest assigning an adult to be an advocate for students who are at-risk. The third recommendation is to make sure the students have academic supports and enrichment through targeted interventions. Fourth, in conjunction with academic supports, students with social and behavioral problems should be identified and supported through targeted interventions to help improve behaviors and functioning. The fifth recommendation is that school wide, students' learning environments and instruction should be personalized to meet their unique educational needs. The final recommendation is to provide rigorous and relevant instruction to help keep students engaged and provide students with skills they will need to graduate and be successful. It should be noted that these recommendations are not backed by "Strong" evidence, and are only supported by research that meets the criteria for "Moderate" or "Low" evidence. Because of this lower threshold of evidence, the authors from the U.S. Department of Education's What Works Clearinghouse suggest that multiple recommendations be implemented as part of a comprehensive approach to preventing high school dropout (Dynarski et al., 2008). As would be expected, the U.S. Department of Education's What Works Clearinghouse's recommendations are consistent with many of the components of the exemplary programs discussed in the review by Hammond et al. (2007). Hammond et al. (2007) also suggest that in order for programs to be more comprehensive and effective, multiple components and strategies that have been shown to be effective should be used. The identified programs and approaches fit into two major categories: (1) dropout prevention, and (2) intervention for students already exhibiting early warning signs for school dropout. The major key components across many of the exemplary programs include having well-trained and qualified staff to implement the prevention or 35 intervention program; take-home resources for students and their parents (i.e., videos, self-help materials, activities, newsletters, and interactive games); a variety of "dosage" levels including length and frequency; and follow-up and booster sessions. Furthermore, the identified programs in the review contained strategies and curriculum focusing on increasing social skills, communication, and problem-solving, as well as targeting academic achievement through homework assistance and tutoring. Many exemplary programs also included a component about helping students to better understand realistic norms for things such as prosocial behaviors, healthy eating habits, sexuality, violence, and substance use. An interactive/role-playing component is typically incorporated when norms are explicitly taught. This strategy has also been shown to be effective in increasing generalization. Another group of researchers conducted a meta-analysis of 152 studies focusing on general dropout programs, and another 15 studies for teen parents (Wilson et al., 2011). The results suggested that both the general and the more specific programs for teen parents were effective. The study further found that programs that had higher levels of implementation quality tended to yield larger effect sizes (Wilson et al., 2011). Interestingly the conclusions drawn from the meta-analysis were that most school- and community-based programs were effective in increasing school completion; therefore, the type of program may not matter, as long as it is being implemented with integrity (Wilson et al., 2011). It is important not only to consider the general conclusions from dropout prevention research, but also specific prevention and intervention programs that are research-based. Check & Connect (Christenson et al., 2012) is based on Finn's (1989) 36 Participation-Identification Model and has shown great outcomes for students. The Check & Connect program uses four tenets: 1) a mentor that builds a longstanding relationship with the student and family, 2) regular checks on student data (academics, behavior, and attendance), 3) timely interventions, and 4) partnership with families. This program is included in the What Works Clearinghouse (Dynarski et al., 2008). The program draws on four different theoretical perspectives: systems-ecological, resilience, cognitive-behavioral, and autonomous motivation. The resilience theory portion of the intervention is the use of a mentor who forms a relationship with the student. The mentors know that it is easier and more effective to draw on the school and community resources than to try to create new resources and programs. The mentors also encourage students to be self-motivated and goal-oriented, and provide many strategies and opportunities for problem solving. This is the autonomous motivation perspective (Reschly & Christenson, 2012). Check & Connect consistently has positive outcomes for students, including improved attendance, improved passing rate, decreased suspensions, and improved dropout rates (Reschly & Christenson, 2012). The program aims to keep students engaged and connected to school. Strategies Specific to the Ninth-Grade Year The transition years from elementary school to middle school and middle school to high school are particularly challenging for students. Students find it challenging to adjust to the new academic and social demands. The experiences students have during these years have a direct impact on student success. Students may begin to disengage without sufficient support and access to a positive school climate (Balfanz et al., 2012). In many school districts, 9th grade is the year many students transition to high school. The 37 9th-grade year is also the year that statistically more students fail then any other school year (National High School Center, 2012). Students who struggle to pass their classes and attend school are then off track for graduation when they are promoted to 10th grade. In many schools, 9th grade is the largest due to students being retained in 9th grade or students beginning to dropout of high school in 10th grade; this phenomenon is known as the 9th-grade bulge and the 10th-grade dip (National High School Center, 2012). For these reasons, the transition years are critical to make sure students remain engaged and connected to school, and on a continued path toward graduation. For the current study, 9th grade was the year students transitioned from middle school to high school; however, in the prior academic year (2013-14) both the 9th-grade and 10th-grade students transitioned at once because of restructuring. The research strongly backs the need to support 1st-year high school students to help prevent a decrease in attendance and grades (Barone, Aguirre-Deandreis, & Trickett, 1991). For example, Reents (2002) found that the schools with transition programs had a dropout rate of 8% compared to 24% at schools without transition programs. However, a survey by the U.S. Department of Education's National Center for Education Statistics, Dropout Prevention Services and Programs in Public Schools and Districts, 2010 to 2011, found that many of the schools sampled did not have adequate transition supports for students. The survey found that only about 40% of districts reported that at least one of their high schools had an advisory period to help students with the transition, only about 26% reported assigning students an adult mentor, and only 20% assigned a student mentor. On a positive note, 77% of the school districts sampled reported the use of one-on-one interventions, in which a school staff member (i.e., counselor, administrator, teacher) provided mentorship 38 to at-risk students; however, these supports were far fewer for smaller and rural school districts. As previously discussed, many school districts rely on adult mentors to provide intervention and support to at-risk students. Research supports having an adult at the school act as a mentor to provide guidance and knowledge to the student (Balfanz et al., 2012); however, the U.S. Department of Education's National Center for Education Statistics, Dropout Prevention Services and Programs in Public Schools and Districts, 2010 to 2011, found that many school districts do not hire additional staff to provide this support to students. Only about 12% of the school districts reported hiring additional staff. School districts reported using community resources more often, at around 30%. This includes community volunteers, child protective services, community mental health agencies, state or local government agencies, churches, or health clinics. Aside from adult mentors, there are many other strategies that are effective in helping students successfully transition from middle to high school. One of the strategies that provides support to transitioning students is 9th grade academies. Ninth-grade academies are learning environments that are either separate from the rest of the school or a completely different school (Reents, 2002). Academies are set up to provide more support to the students, and help to make the transition less overwhelming. Another similar approach keeps students in the same small learning academies for two to four years; groupings that are based on students' interests are called career academies (Brand, 2009). Herlihy (2007) from the National High School Center identified five ways schools can help ensure students successfully transition from 9th to 10th grade. First, the school 39 should have an established data monitoring system, much like an EWS that was discussed previously, where school personnel can easily identify at-risk students. Second, the school needs to consider the students' instructional needs, and make sure students are receiving appropriate curricular supports and classes. Third and fourth, schools need to personalize student learning to address individual needs; to do so, the school should have a wide range of supports and services available to help personalize a student's learning environment. Lastly, it is important for schools to help students make the connection of why their education is important, and how it can be necessary for future employment and/or admissions into colleges/universities. Barber and Olsen (2004) found that students in 9th grade perceive the supports and activities available to them differently than students in other grades. Ninth-grade students' perception is that they have less support from teachers and principals then they did in middle school. In general, they also report liking school less. Students in 9th grade also report being less involved in school activities, but conversely, say that there need to be more school organizations. These students also reported lower self-esteem and higher rates of depression than middle school students. The Link Crew program is an evidence-based transition program. Link Crew aims to help 9th-grade students feel welcome and connected in their 1st year of high school. The program trains 11th- and 12th-grade students to be peer mentors to 9th-grade students to help them have a successful freshman year. Link Crew spans the entire freshman year, and includes a high school orientation, academic follow-ups, social follow-ups, and leader initiated contacts. There is also a school safety and antibullying component built in (Boomerang Project, 2011). Research suggests that schools that implemented the Link 40 Crew program saw improvements in office referrals with a 37% decline, suspensions with a 20% decline, absences with a 33% decline, and tardies with a 7% decline (Boomerang Analysis, 2011). There was also a 6% decrease in Ds and Fs among 9th-graders, and a 3% drop in the number of students who had failed one class or failed more than three classes (Boomerang Analysis, 2011). The Impact of Extracurricular Activities Since the 1960s, researchers in a number of different fields (sociology, economics, and psychology) have studied the impact of extracurricular activities on the development of children and adolescents (Feldman Farb & Matjasko, 2012). The impact of extracurricular activities is often explained using Bronfenbrenner's (1979) ecological theory. An adolescent's participation in an extracurricular activity helps shape their environment through their interactions with peers, mentors/coaches, and the characteristics of the activity. These interactions then in turn influence their development. Feldman Farb and Matjasko (2012) recently reviewed research on the study of extracurricular activities and the impact on adolescent development. They also summarized results from a review by Holland and Andre (1987) and their previous review (Feldman & Matjasko, 2005). Holland and Andre (1987) found that most studies demonstrated a positive relationship between participation in athletics and adolescent development. Their previous review found that the research since Holland and Andre's (1987) review was mainly replication studies with an expansion from solely looking at athletics to now looking at other extracurricular activities. Analysis techniques were also updated (Feldman & Matjasko, 2005). The results from their previous review indicated that the studies had mixed results. For example, some studies found that there were 41 negative outcomes to participation in extracurricular activities such as sports because participation in sports was correlated with higher substance use and sexual activity in males (Feldman & Matjasko, 2005); however, many of the studies found that participation in extracurricular activities was generally positive for students' development. The most current review addressed a number of limitations in previous research including: measurement of participation; intensity, breadth, and duration of participation; person-centered approaches; and threshold effects (Feldman Farb & Matjasko, 2012). These advancements helped to further differentiate outcomes for groups and helped to define when extracurricular participation is no longer beneficial. The review found many of the same positive associations. For example, they found that continued participation was associated with attainment of educational goals, college acceptance, and prosocial behaviors. Further research suggested that there was a threshold and that past a certain point there was no further relationship with positive outcomes. This is known as the overscheduling hypothesis (Feldman Farb & Matjasko, 2012). The review cited a study by Luther and colleagues (2006) that used cluster analysis to identify groups of students among an affluent student sample who were overscheduled. They did not find any particular group of students that met criteria for being overscheduled, but they did find that student perceptions about parental pressure and lack of afterschool supervision led to poorer adjustment. This suggests that there could be other moderating variables that could explain differences in outcomes among students. Two other studies explored the relationship between overscheduling and psychological adjustment, and found a relationship between time spent in the activity and anxiety levels (Melman, Little, & Akin Little, 2007). Another study completed by 42 Mahoney and colleagues (2006), found that Black youth who participated in extracurricular activities for more than 20 hours had lower self-esteem, suggesting that this could possibly be due to spending less time having meaningful discussions with their parents. The study also found that across groups, youths who spent 15 or more hours a week in organized activities, were more likely to report more alcohol use than youth who spent less than 15 hours in extracurricular activities. All of the studies found somewhat different results, suggesting that the number of hours spent in an activity does not necessarily equate to better outcomes and additional variables need to be considered. The review found that studies that attended to the intensity, breadth, and duration of participation in extracurricular activities identified qualitative differences in adolescent experiences that could potentially impact developmental outcomes (Feldman Farb & Matjasko, 2012). In general, the way researchers measure intensity is by looking at the number of hours of participation a week or the number of days per week participating in the activity. The review found that there is typically a positive relationship between academic grades, long-term aspirational goals, and educational goal attainment (Darling, 2005; Denault & Poulin, 2009). Breadth, or the sum of different types of activities, was another variable that was analyzed. In general, there was a positive relationship between the number of different activities an individual is involved in and the individual's academic and other school outcomes; however, there was a threshold effect with students not having continued positive effects if participating in more than six to eight different activities (Rose-Krasnor et al., 2006). Breadth was usually the variable that explained variance in outcomes, suggesting that adolescents who are involved in a variety of activities are typically more well-adjusted (Feldman Farb & Matjasko, 2012). 43 The review also reported results from two studies that looked at different types of group involvement; this approach is called person-centered (Feldman Farb & Matjasko, 2012). The person-centered approach uses cluster analysis to categorize different types of participation styles. From the person-centered research, there were two very interesting findings regarding student engagement. First, students who were categorized as "unstructured," who spent their time playing video games or engaging in nonschool related activities, showed the poorest school perception, poorest attendance, and lowest academic achievement in comparison to other groups. In contrast, students who were categorized into the "all-around" group, who participated in a number of different activities (school-based and nonschool-based), showed the best outcomes for school perception, attendance, and academic achievement (Nelson & Gastic, 2009). Secondly, similar results were found in at-risk adolescent samples (Metzger, Crean, & Forbes-Jones, 2009; Peck et al., 2008). The research using person-centered approaches is a promising avenue because it sheds light on qualitative differences in how different participation styles impact achievement outcomes. In general, the extracurricular activity research suggests that participation in school-based extracurricular activities generally has a positive impact on adolescent development and student outcomes. The research also indicates, however, that it is important to consider group differences, duration, frequency, intensity, and threshold effects as potential moderators for student and adjustment outcomes. Utah Dropout Trends and Prevention Initiatives There are trends nationally and locally to focus efforts on increasing high school graduates' knowledge and skills in order to be ready for college and/or a career. In Utah, 44 the Strengthening the Senior Year/Career and College Ready (SY/CCR) Work Group was created by the Utah State Office of Education (USOE) to address the issues concerning college and career readiness. The SY/CCR Work Group evaluated effective and innovative practices that have worked at the state and national levels, and found three common themes on which they based their recommendations: (1) "providing rigorous and relevant coursework;" (2) "connecting students with multiple pathways, options, and supports;" and (3) "strengthening education and career planning by providing, effective guidance and planning systems and processes at every level from grade six through grade 12" (Utah State Office of Education, 2010). The Work Group's report states that high-stakes assessments that potentially affect college and career decisions are completed prior to senior year, which provides little incentive for students to try hard during one's senior year. In fact, students who may not have done well on previous high-stakes assessments and struggled academically may feel like giving up academically and possibly dropping out (Utah State Office of Education, 2010). To better understand high school completion rates in Utah, the USOE began tracking students from their 10th-grade year to the end of their 12th-grade year in 2007 (Utah State Office of Education, 2010). Their findings suggest that the graduation rate in Utah is around 88%, and has been consistent over the past 3 years; however, there are disparities in graduation rates across different ethnic and disadvantaged groups. Students who were identified as having a disability had a graduation rate of 81%, economically disadvantaged students had a graduation rate of 78%, African American students' graduation rate was 77%, American Indian students' was 74%, Hispanic students' was 45 71%, and English Learners (ELs) was 69% (Utah State Office of Education, 2010). The most recent data from the Utah State Office of Education (2013) suggest that the graduation rate for the 2013 graduating class was 81%, calculated using the 4-year cohort rate. Graduation rates have increased by 3 percentage points from 2012 to 2013. The recent data also suggest that the graduation rate for Hispanic students has increased by 5 percentage points, for students identified with a disability 4 percentage points, and for ELLs 9 percentage points, from 2012 to 2013 (Utah State Office of Education, 2013). Although graduation rates are increasing for many subgroups, these data continue to indicate that the needs of a large number of students in the Utah public school system are not being met, and that there is a need for system improvement. The USOE's findings and research suggest that there is a need for curriculum to be relevant and engaging for students. With that said, students have differences in what is relevant and interesting to them; all students should feel that they are able to access an education that meets their own educational and/or career goals. There are many programs being implemented in Utah schools that potentially meet the varying educational/career needs of students, including work-based learning, career pathways, alternative education, acceleration and enrichment opportunities, youth options, and collaborative initiatives to improve attendance (Utah State Office of Education, 2010). The main purpose of these programs is to keep students engaged, and continue to provide meaningful learning opportunities. The USOE (2014) has also created a Utah Statewide Dropout Prevention Committee to "identify a set of strategies and practices that are key components of interventions that have demonstrated promise in reducing dropout rates" (p. 9). The 46 committee came up with a practice guide and the recommendations within the practice guide are made up of a combination of what Local Education Agencies (LEAs) are currently using across the state that have shown good outcomes and strategies promoted by the IES Guide for Dropout Prevention (Dynarski et al., 2008; USOE, 2014). The practice guide consists of six recommendations: using data systems that report the number of students who drop out and identify students at-risk for dropping out, assigning adult mentors to at-risk students, providing targeted academic interventions to at-risk students, implementing targeted behavior and social skills interventions, personalizing the learning environment and instruction, and providing rigorous and relevant instruction to engage all students and provide them with skills that they can use in their postsecondary schooling and/or career (Utah State Office of Education, 2014). These recommendations are consistent with guidelines set by the U.S. Department of Education's What Works Clearinghouse (Dynarski et al., 2008). Rationale for Current Study High school dropout prevention continues to be an issue in the United States' educational system with a graduation rate of around 80.0% (Stetser & Stillwell, 2014). This is the lowest student dropout numbers have ever been, partly because schools continue to try to improve in their efforts to intervene with students who may be at-risk. Much of the research on dropout prevention has focused on identifying risk factors, and there are a number of factors that potentially place a student at risk for dropout. Of those, there are a few factors that have been pinpointed as predominant risk factors and red flags for student disengagement: attendance, behavior, and academic performance, especially in core classes. 47 EWSs have been recommended as a way to track attendance, behavior, and academic concerns. Schools can easily utilize an EWS framework to help identify those students who may be at-risk and in need of additional support and intervention. Furthermore, increased access to electronic student data has helped improve the ability of schools to track those students who may be at-risk. These students can then be provided with appropriate interventions and supports after they have been identified as at-risk and in need of services. It is important for schools to make sure that the data from the EWS are used efficiently to identify and place students in intervention services. Further challenges include that many schools have limited resources to meet the needs of all of their at-risk students, and many school programs have not been assessed for efficacy. Therefore, it is important to better understand how current school-based intervention and prevention programs can be effective for at-risk students who have been identified through EWSs. The current study aimed to assess whether the target school's EWS and prevention/intervention programs were effective in improving outcomes of attendance, behavior, and academic performance. Student disengagement from school is another major risk factor that researchers and school personnel can measure, monitor, and change. Appleton et al. (2006) conceptualized student engagement into four categories: academic, behavioral, psychological, and cognitive. The current study will focus on examining the impact of psychological and cognitive engagement on student outcomes. Appleton and Christenson (2004) created a student self-report measure to more accurately measure psychological and cognitive engagement, called the SEI. The current study used the SEI to assess students' psychological and cognitive engagement, while the EWS was used to track 48 attendance, academic performance, and reported behavior problems as indicators of dropout risk. Furthermore, the study aimed to look at how participation in school-based at-risk programs and extracurricular activities potentially act as protective factors for student engagement. This is important because few studies have looked at the impact of school-based dropout prevention programs on student engagement and the EWI variables of attendance, behavior, and academic performance while also evaluating the effectiveness of the EWS. The study also attempted to answer secondary research questions focused on the 9th-grade transition year. The target high school implemented a universal transition program for the incoming 9th-grade class. The transition program used assigned peer mentors, who were 11th- and 12th-grade students, to help provide support and insight to 9th-grade students. Since this was a new program the past school year, the previous year's 9th-grade class did not have assigned peer mentors. To assess the effectiveness of the transition program, the study examined differences in EWIs and student engagement between the 9th-grade students and 10th-grade students (the previous year's 9th graders). The study also examined improvement in EWIs and student engagement variables from fall to the spring for both 9th- and 10th-grade students. Also, since previous research has found differences in dropout rates based on income, race/ethnicity, and other factors, the study included analyses of the influence of these demographic variables on student outcomes. Research Questions The following questions were the focus of this research project: 1. Are students who are identified through an Early Warning System (EWS) as "at- 49 risk," the same students who are connected to school-based supports and at-risk services? 2. Does participation in school-based supports and at-risk programs result in an increase in students' self-report of cognitive and psychological engagement? a. Does participation in school-based at-risk programs increase students' psychological engagement on the Student-Teacher Relationships factor? b. Does participation in school-based at-risk programs increase students' psychological engagement on the Peer Support for Learning factor? c. Does participation in school-based at-risk programs increase students' psychological engagement on the Family Support for Learning factor? d. Does participation in school-based at-risk programs increase students' cognitive engagement on the Future Aspirations and Goals factor? e. Does participation in school-based at-risk programs increase students' cognitive engagement on the Control/Relevance factor? 3. Is there a correlation between students' self-reported level of cognitive and psychological engagement and specific Early Warning Indicator (EWI) variables? a. Is there a positive correlation between students' cognitive and psychological engagement and GPA? b. Is there a negative correlation between students' cognitive and psychological engagement and days absent? c. Is there a negative correlation between students' cognitive and psychological engagement and discipline referrals? 4. Does participation in school-based supports and at-risk programs correlate with 48 attendance, academic performance, and reported behavior problems as indicators of dropout risk. Furthermore, the study aimed to look at how participation in school-based at-risk programs and extracurricular activities potentially act as protective factors for student engagement. This is important because few studies have looked at the impact of school-based dropout prevention programs on student engagement and the EWI variables of attendance, behavior, and academic performance while also evaluating the effectiveness of the EWS. The study also attempted to answer secondary research questions focused on the 9th-grade transition year. The target high school implemented a universal transition program for the incoming 9th-grade class. The transition program used assigned peer mentors, who were 11th- and 12th-grade students, to help provide support and insight to 9th-grade students. Since this was a new program the past school year, the previous year's 9th-grade class did not have assigned peer mentors. To assess the effectiveness of the transition program, the study examined differences in EWIs and student engagement between the 9th-grade students and 10th-grade students (the previous year's 9th graders). The study also examined improvement in EWIs and student engagement variables from fall to the spring for both 9th- and 10th-grade students. Also, since previous research has found differences in dropout rates based on income, race/ethnicity, and other factors, the study included analyses of the influence of these demographic variables on student outcomes. Research Questions The following questions were the focus of this research project: 1. Are students who are identified through an Early Warning System (EWS) as "at- 51 students' psychological engagement on the Family Support for Learning factor? d. Does participation in the 9th-grade student transition program increase students' cognitive engagement on the Future Aspirations and Goals factor? e. Does participation in the 9th-grade student transition program increase students' cognitive engagement on the Control/Relevance factor? 7. Is there a difference in Early Warning Indicator (EWI) variables between the 9th-grade class who participated in the transition program and the 10th-grade class who did not participate in the program the previous school year? a. Does the 9th-grade class have higher GPAs than the 10th-grade class? b. Does the 9th-grade class have fewer days absent than the 10th-grade class? c. Does the 9th-grade class have fewer discipline referrals than the 10th-grade class? 8. Do differences in demographic background variables of socioeconomic status, race, ELL status, gender, grade, and/or middle school of origin correlate with participation in at-risk programs, level of cognitive and psychological engagement, and/or Early Warning Indicator (EWI) outcome variables? CHAPTER 2 METHOD Participants The participants were enrolled in a high school in a suburban school district within a large Western city. The high school had around 2000 students enrolled throughout the school year. The Student Engagement Inventory (SEI) was administered to 9th-, 10th-, 11th-, and 12th- grade students attending the target high school in the spring of the 2014-15 school year to gather information about students' cognitive and psychological engagement. The spring data collection had an initial sample size of 1,467, which was around 75% of the entire student population; this represents the number of students who completed the Student Engagement Inventory (SEI). There were some missing data points on important variables such as the Student Engagement Instrument (SEI), but the missing cases on each item the SEI were fairly small in comparison to the sample size (0.4 % to 2.1%). The missing cases were removed because there was such a small percentage compared to the sample, and a large sample size would remain for analysis. The remaining sample size was N=1,314. This sample was referred to as Sample 1 (see Table 1). The sample was 52.7% (n=693) male and 47.3% (n=621) female. The sample was made up of 25.0% (n=328) 9th graders, 28.8% (n=378) 10th graders, 25.6% 11th graders (n=336), and 20.7% (n=272) 12th graders. The self-reported race make-up of the school sample was as follows: White 71.9% (n=945), Hispanic-Latino 9.9% (n=130), 53 American Indian 0.5% (n=6), Asian 1.9% (n=25), African American/Black 2.0% (n=26), Pacific Islander 1.9% (n=25), Multiracial 9.8% (n=129), and Other 2.1% (n=28). Since many of the racial groups had small sample sizes, the data were broken into the groups "White" (71.9%, n=945) and "Non-White" (28.1%, n=369). The proportion of students in special education programs was 8.2% (n=108) and 4% (n=57) of students were identified as English Language Learners. The information on low income students was based on the list of students who qualified for fee waiver and free and reduced lunch. Unfortunately, the list available from the school was missing many of the students who participated in the survey and only 916 students were matched to the sample. Of those students, 27.1% (n=248) were low income students, and 72.9% (n=668) were not low income students. Based on risk level, Sample 1 included 50.3% (n=661) "low risk" 28.1% (n=369) "at-risk" 21.6% (n=284) "significant risk" students. The sample was also analyzed for participation in at-risk and prevention programs. In the at-risk programs, 1.8% (n=23) students participated in the Reading Class, 7.2% (n=94) participated in Math Lab, 5.5% (n=72) participated in the Study Skills class, and 8.2% (n=108) received special education classes and support services. In the prevention programs, 3.9% (n=51) participated in Latinos in Action (LIA), and 8.4% (n=111) participated in Advancement via Individual Determination (AVID). The SEI was also previously administered just to 9th and 10th graders in late January 2015 to assess change in student engagement over time (January to Spring). Data from the two administrations were combined to help answer some of the secondary research questions. The total sample size for 9th- and 10th-grade students who took the SEI during both administrations was 746, which was around 75% participation rate of the 54 9th- and 10th-grade student population. Missing data cases were removed and the final sample size for 9th- and 10th-grade students with two complete data points was 596. This sample was referred to as Sample 2. There were 318 (53.4%) male students and 278 (46.6%) female students. The sample had 44.1% (n=263) 9th-grade students and 55.9% (n=333) 10th-grade students. The self-reported racial make-up was as follows: White 68.0% (n=405), Hispanic 12.9% (n=77), Black 2.3% (n=14), Asian 1.7% (n=10), Pacific Islander 1.8% (n=11), American Indian 0.7% (n=4), Other 2.3% (n=14), and Multiracial 10.2% (n=61). The proportion of students in special education programs was 8.1% (n=48) and there were 5% (n=32) of students who were identified as English Language Learners (ELL). The information on low income students was based on the list of students who qualified for fee waiver and free and reduced lunch. Unfortunately, the list was missing many of the students who participated in the survey and only n=301 students were matched to the sample. Of those students, 33.2% (n=100) were low income students and 66.8% (n=201) were not low income students. The demographic information for Sample 2 is in Table 2. Setting The participating high school was located in a large suburban school district, and served students from a wide range of socioeconomic backgrounds. The school district has an electronic Early Warning System (EWS) that tracks the EWI variables of attendance, grades, and behavior. The EWS data are used by problem-solving teams to help identify students as at-risk and connect those students to appropriate school supports and programs. The school offers a plethora of extracurricular school-based activities for students, including several at-risk programs. 55 The at-risk programs in place at the school that were included for analysis in the current study are Advancement via Individual Determination (AVID), Latinos in Action (LIA), Study Skills, Math Lab, Reading class, and Special Education. AVID was initially created for students in grades 6 through 12, who are typically the underrepresented "academic middle" students, and provides structured teaching methods to help make the curricula more accessible to the students. The staff are highly trained to help implement AVID goals. Further, the school staff aim to eliminate low-level tracking and provide academic and motivational supports (San Diego County Office of Education, 1991). Adult tutors support the teachers, and help lead Socratic seminars in classrooms. There are research studies that support the effectiveness of AVID as a school reform model to improve student outcomes. Watts and colleagues (2002) found that students who were in the AVID program preformed better on standardized testing and attended school more often. A 4-year longitudinal study by Guthrie and Guthrie (2000), found that students who were in the AVID program for 2 years had higher GPAs than those with only 1 year or no AVID experience, earned credits that placed them on track for 4-year college acceptance, and took more Advanced Placement (AP) classes than students with 1 year or no AVID experience. They found that 95% of the AVID graduates were enrolled in a college or university following high school (Guthrie & Guthrie, 2000). AVID is included on the Hammond et al. (2007) Exemplary Program list of programs with high levels of evidence supporting its efficacy as an effective program in preventing high school dropout. LIA is a program that utilizes a classroom format to help empower Latino students in middle and high school through culture, service, and academic achievement. 56 Enriquez (2012) found that Latino students enrolled in LIA reported higher levels of school engagement, desire for educational attainment, and feelings that school was a major factor in self-understanding than their Latino peers not enrolled in LIA. The study also found that the students involved in LIA increased their leadership and social skills, and their drive for school success (Enriquez, 2012). The participating school also had a Study Skills class, which provided students with explicit instruction on study skills necessary to gain access to the curriculum, and become effective and independent learners. Paulsen and Sayeski (2013) emphasize that successful high school students have effective management skills such as, study habits, time management, and self-management, and cognitive study habits including, interpreting visuals, using references, and taking notes. Explicit study skills instruction and mentoring of students about academics and school attendance are components in exemplary programs (Hammond et al., 2007). The effectiveness of the Study Skills program offered by the school was not assessed previously, but does include efficacious practices as reported in the research literature. Aside from the Study Skills program, the participating high school also provided a Math Lab (math class) and a Reading class for students with low math and/or reading achievement. Math Lab could also be used as credit recovery for students who may have failed a math class. Math Lab is a form of "double-dosing" in which students' receive more instructional time, which gives them more opportunities to learn and retain the curriculum (Cortes, Goodman, & Nomi, 2013). There is evidence that double dosing has a positive and long-lasting effect on student achievement in math (Cortes et al., 2013). Math Lab also preteaches concepts, which has been shown to be effective in many 57 academic areas including math. Lalley and Miller (2006) specifically looked at differences between preteaching and reteaching math concepts, problems, and computation. They found that both effectively increased knowledge of math concepts, math computation, and problem mastery. The study also found that students who were in the preteaching group significantly increased their self-concept related to math abilities, where the reteaching group did not. Improvements in self-concept could potentially impact school engagement. The Reading class is designed for students who are struggling with reading fluency or decoding. It is offered to students who are identified through a double-gating procedure as at-risk for academic failure due to reading difficulties. Students are first administered the Scholastic Reading Inventory (SRI), and then administered a Curriculum-Based Measurement (CBM) of reading fluency. Students who qualify are strongly encouraged to enroll in the reading class in order to improve their likelihood of success in their high school classes. The participating school also had a school-wide peer mentoring program for 9th-grade students transitioning to high school, called Link Crew, which aimed to help 9th-grade students feel welcome and connected in their 1st year of high school. The program trains 11th- and 12th- grade students to be peer mentors to 9th-grade students to help them have a successful freshman year. Link Crew spans the entire freshman year, and includes a high school orientation, academic follow-ups, social follow-ups, and leader initiated contacts. There is also a school safety and antibullying component built in (Boomerang Project, 2011). Research suggests that schools that implemented the Link Crew program saw improvements in office referrals with a 37% decline, suspensions with a 20% 58 decline, absences with a 33% decline, and tardies with a 7% decline (Boomerang Analysis, 2011). There was also a 6% decrease in D and F grades among 9th-graders, and a 3% drop in the number of students who had failed one class or failed more than three classes (Boomerang Analysis, 2011). This was the 1st year the school used the Link Crew program. The Link Crew student leaders were trained for 20 hours during the summer and participated in freshman orientation. The Link Crew Leaders were trained to spend around 2 hours per month with their assigned students throughout the school year. It was reported that some peer leaders went above and beyond that requirement, while other peer leaders rarely met with their students. To better address these compliance issues and improve the support to freshman students, the program will reportedly expand in coming years and offer a class to students who are dedicated to being Link Crew Leaders. This class will provide more time to cultivate a positive culture, create leaders who can have a positive effect on students, and help the positive culture spread throughout the school. Students receiving special education services were included in the sample of the current study. Special education is instruction and services that are specialized to meet the unique needs of a child with a disability (National Dissemination Center for Children with Disabilities, 2013). Services can range from a self-contained classroom/special school to only receiving special education services with a related service provider, such as an occupational therapist, school psychologist, or speech and language pathologist. Special education was created to ensure that individuals with disabilities received a free and appropriate education (FAPE) (U.S. Department of Education, 2004). Students who received special education services were included in the study because many students with a learning or social/emotional disability are at greater risk for dropout and school 59 disengagement (Deshler et al., 2001; Reschly & Christenson, 2006; Sabornie & deBettencourt, 2004). Furthermore, individuals with learning and social/emotional disabilities are more likely to have multiple risk factors (Wagner et al., 2006). Students who received special education services solely in the two self-contained special education classes due to significant intellectual disabilities and low reading abilities were excluded from the study because they did not have the skills necessary to comprehend the written study measures. Students receiving special education services whose skill level also permitted inclusion in the regular curriculum in history or English classes were included in the administration of the study measures. Measures Independent Variables Participation Variables Student Survey A student survey was used to obtain self-report information regarding student participation in any extracurricular activities, at-risk, and special education programs. In addition to the number of different programs and activities students were involved in, students also reported on the duration of their participation in these activities. The Student Survey also collected demographic data and information on students' perception of adult mentors and other school-based supports in the school. Observed Variables The observed variables associated with the latent variable of Participation are any special education classes or other related services that students received, including AVID, 60 LIA, Study Skills, Math Lab, and/or Reading Class. These variables were treated as dichotomous variables. Students also reported the total number of hours they participated in either an extracurricular activity or at-risk/prevention program. The variable of time was going to be used if the dichotomous variables did not work well within the model. The time variable was treated as a continuous variable. It should be noted that the entire 9th-grade class participated in the transition program, Link Crew; therefore, the secondary model considered 9th graders as a treatment group. Grade was treated as a dichotomous variable. Demographic Variables Participating students' gender, race/ethnicity, language spoken in the home, what middle school they attended, and socioeconomic status (SES) based on fee-waiver eligibility was collected using Skyward (1999 to present) and Data Dashboard. Skyward and Data Dashboard are electronic-based systems that the target school uses regularly to access student information. Student data were grouped by demographic information and the models were tested for group differences. Dependent Variables Student Engagement Variables All students were assessed for their perceived level of school engagement during the Spring Semester (April) during the 2014-2015 school year using the Student Engagement Instrument (SEI; Appleton & Christenson, 2004). The spring administration was used to answer questions for the school-wide model. The perceived level of student engagement was also assessed in late January for 9th- and 10th-grade students. The 61 January and the spring administrations were used to answer the secondary research questions regarding the impact of the transition program by comparing 9th- and 10th-grade student data. Student Engagement Instrument (SEI) The Student Engagement Instrument (SEI) assesses a student's perceived level of engagement. The scale includes 35 items and six subscales. The six subscales are organized under two domains of engagement: cognitive engagement and psychological engagement. Cognitive engagement is self-regulation, being able to connect schoolwork to future goals, valuing learning, setting personal goals, and autonomy. The three subscales that load onto the cognitive engagement domain are: (1) control and relevance of schoolwork, (2) future goals and aspirations, and (3) extrinsic motivation. It should be noted, however, that the extrinsic motivation subscale was removed from the model analyses because it is only comprised of two items and can be problematic in data analyses (Appleton et al., 2006). Psychological engagement includes feelings of belongingness or identification with school and peers, and relationships with teachers and peers. It refers to engagement that is interpersonal in nature. The three subscales that load onto the psychological engagement domain are: (1) teacher-student relationships, (2) peer support for learning, and (3) family support for learning. Dr. Angie Pohl of the Check and Connect team at the Institute on Community Integration, University of Minnesota was consulted about the use of the SEI, as there is no accompanying manual; however, there are research articles outlining standardization and validation methods, as well as how to use the instrument. It should be noted that the researcher requested a Spanish version for ELL students, but there was not one available. Native Spanish speakers were used to 62 accurately translate the document to Spanish. A study by Appleton et al. (2006) aimed to validate the SEI. The instrument was normed in a large, diverse, urban school district in the Midwest. The study included 9th-grade students, and of the 2,577 students selected for the study, 1,931 completed the SEI, which was about a 75% participation rate. The ethnic make-up of the sample was 40.4% African American (n=780), 35.1% White (n=677), 10.8% Asian (n=208), 10.3% Hispanic (n=199), and 3.5% American Indian (n=67). The gender numbers were about equal with girls making up 51% of the sample. There were 22.9% of the students who reported speaking another language in the home other than English, 61.4% qualified for free and reduced lunch, and 7.6% of the sample received services through special education. Appleton et al. (2006) used exploratory factor analysis and confirmatory factor analysis to find the best model fit. The validation study supported the six factors, but also reported that the extrinsic motivation factor should be further researched to better understand if it should be included in the model. The extrinsic motivation factor was removed for the current study, and a five-factor model was used. The factors correlated with expected educational outcomes. Risk Variables The study examined Early Warning Indicators (EWIs) of attendance, behavior, and academic records as the main at-risk indicators. Each EWI loaded onto the latent variable of Risk. Similar information was collected monthly throughout the study to assess for change over time, and correlations with engagement variables; however, the EWIs from the end of the school year were used to answer the primary research questions. Thi